Nut upsuser - Sep 2012 - Triplite Lock's when on battery

Hey all,

I have a triplite 1050 and when I did a test today  I saw that the 
battery bounce from 100 to 80 to 90 to 91 percent before UPSD lost its 
connection to the UPS. The server then decided that it was time to shut 
off.

The UPS sat at 91% for about 5 min before UPSD lost its connection, so 
I'm not sure if it wasn't updating its battery.charge... I wasn't
sure
if this is expected behavior...

Luckily this was a planned test and my script just outputs to messages 
that it believes it is time to shut off.


upsc ups
battery.charge:  91
battery.test.status: Battery OK
battery.voltage: 13.00
battery.voltage.nominal: 18
driver.name: tripplite_usb
driver.parameter.pollinterval: 2
driver.parameter.port: auto
driver.version: 2.2.2
driver.version.internal: 0.11.1
input.voltage: 0.00
input.voltage.nominal: 120
output.voltage: 118.0
ups.debug.load_banks: 0
ups.debug.V: 31 30 33 30 58 58 0d '1030XX.'
ups.delay.shutdown: 64
ups.firmware: F0957.B
ups.firmware.aux: protocol 2001
ups.mfr: Tripp Lite
ups.model: OMNISMART1050
ups.power.nominal: 1050
ups.status: OB


Sep 25 15:34:43 server kernel: hiddev96: USB HID v1.00 Device [TRIPP 
LITE TRIPP LITE OMNISMART1050      ] on usb-0000:00:1f.2-2

Broadcast message from nut (Tue Sep 25 15:34:44 2012):

Communications with UPS ups at 127.0.0.1 lost
Sep 25 15:34:44 server upsmon[18224]: Poll UPS [ups at 127.0.0.1] failed - 
Driver not connected
Sep 25 15:34:44 server upsmon[18224]: Communications with UPS 
ups at 127.0.0.1 lost
Sep 25 15:34:44 server wall[3925]: wall: user nut broadcasted 1 lines 
(44 chars)
Sep 25 15:34:49 server upsmon[18224]: Poll UPS [ups at 127.0.0.1] failed - 
Driver not connected

Broadcast message from nut (Tue Sep 25 15:34:59 2012):

Executing automatic power-fail shutdown

On Sep 25, 2012, at 4:31 PM, twhitney at clearbearing.com wrote:
> Hey all,
> 
> I have a triplite 1050 and when I did a test today  I saw that the battery
bounce from 100 to 80 to 90 to 91 percent before UPSD lost its connection to the
UPS. The server then decided that it was time to shut off.
> 
> The UPS sat at 91% for about 5 min before UPSD lost its connection, so
I'm not sure if it wasn't updating its battery.charge... I wasn't
sure if this is expected behavior...
If upsd loses its connection to the driver, and the last status was "on
battery", it assumes the worst, and attempts to shut down.
> Luckily this was a planned test and my script just outputs to messages that
it believes it is time to shut off.
> 
> 
> upsc ups
> battery.charge:  91
> battery.test.status: Battery OK
> battery.voltage: 13.00
> battery.voltage.nominal: 18
> driver.name: tripplite_usb
> driver.parameter.pollinterval: 2
> driver.parameter.port: auto
> driver.version: 2.2.2
> driver.version.internal: 0.11.1
> input.voltage: 0.00
> input.voltage.nominal: 120
> output.voltage: 118.0
> ups.debug.load_banks: 0
> ups.debug.V: 31 30 33 30 58 58 0d '1030XX.'
> ups.delay.shutdown: 64
> ups.firmware: F0957.B
> ups.firmware.aux: protocol 2001
> ups.mfr: Tripp Lite
> ups.model: OMNISMART1050
> ups.power.nominal: 1050
> ups.status: OB
> 
> 
> Sep 25 15:34:43 server kernel: hiddev96: USB HID v1.00 Device [TRIPP LITE
TRIPP LITE OMNISMART1050      ] on usb-0000:00:1f.2-2
> 
> Broadcast message from nut (Tue Sep 25 15:34:44 2012):
> 
> Communications with UPS ups at 127.0.0.1 lost
> Sep 25 15:34:44 server upsmon[18224]: Poll UPS [ups at 127.0.0.1] failed -
Driver not connected
> Sep 25 15:34:44 server upsmon[18224]: Communications with UPS ups at
127.0.0.1 lost
> Sep 25 15:34:44 server wall[3925]: wall: user nut broadcasted 1 lines (44
chars)
> Sep 25 15:34:49 server upsmon[18224]: Poll UPS [ups at 127.0.0.1] failed -
Driver not connected

It looks like the kernel rediscovered the UPS at 15:34:43. How much time had
elapsed since you cut power?

This version of NUT (2.2.2) is fairly old - newer versions of the USB code in
NUT will attempt to reconnect more aggressively. I don't remember if this
particular version is prone to crashing when an UPS disconnects, but that might
have been the impetus behind some of the changes since then.

Also, is the UPS plugged directly into the server with a short USB cable? UPS
tests and power failures can cause USB interference. A hub should theoretically
work, but it is probably not something to rely on.

-- 
Charles Lepple
clepple at gmail

On 09/29/2012 04:36 AM, Charles Lepple wrote:
> What if you just comment out "autoreconf -i"? It really
shouldn't be
> necessary. If the generated ./configure script is wrong, chances are 
> we can fix it in the NUT source (unless RHEL is even more busted than 
> I thought). 
     That did it but now I have a few problems with some of the drivers

     I have tried to disable the freeipmi because it fails (despite 
having freeipmi-devel included as a build requirement):

    checking for FreeIPMI version via pkg-config... none found
    checking for FreeIPMI cflags...
    checking for FreeIPMI ldflags... -lfreeipmi -lipmimonitoring
    checking for freeipmi/freeipmi.h... yes
    checking for ipmi_monitoring.h... yes
    checking for library containing ipmi_ctx_create... no
    checking for library containing ipmi_ctx_find_inband... no
    checking for library containing ipmi_fru_parse_ctx_create... no
    checking for library containing ipmi_monitoring_init... none required
    checking for library containing
    ipmi_monitoring_sensor_read_record_id... no
    checking for library containing ipmi_sdr_cache_ctx_destroy... none
    required
    checking for library containing ipmi_sdr_ctx_destroy... no
    configure: error: "FreeIPMI not found, required for IPMI support"

      I assumed that the version included in RHEL5 might be too old, so 
what I did was:

%configure \
     --with-all \
     --without-powerman \
     --without-freeipmi \
[...]


but it still tries (and fails) to detect the ipmi libs, after which it 
bails out:

    checking whether to enable Avahi support... yes
    checking whether to build Powerman PDU client driver... no
    checking for FreeIPMI version via pkg-config... none found
    checking for FreeIPMI cflags...
    checking for FreeIPMI ldflags... -lfreeipmi -lipmimonitoring
    checking for freeipmi/freeipmi.h... no
    checking for ipmi_monitoring.h... no
    checking for library containing ipmi_ctx_create... no
    checking for library containing ipmi_ctx_find_inband... no
    checking for library containing ipmi_fru_parse_ctx_create... no
    checking for library containing ipmi_monitoring_init... no
    checking for library containing
    ipmi_monitoring_sensor_read_record_id... no
    checking for library containing ipmi_sdr_cache_ctx_destroy... no
    checking for library containing ipmi_sdr_ctx_destroy... no
    configure: error: "FreeIPMI not found, required for IPMI support"


     Do you have any quick idea ?

     manuel
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.alioth.debian.org/pipermail/nut-upsuser/attachments/20120929/21559e2f/attachment-0001.html>